Poverty and Shared prosperity in...

POVERTY AND SHARED PROSPERITY IN COLOMBIA

BACKGROUND PAPER FOR POLICY NOTES

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

Pub

lic D

iscl

osur

e A

utho

rized

1

Contents I. Poverty, Inequality and Shared Prosperity ........................................................................................... 2

Poverty: recent trends .................................................................................................................................. 2

Profiling the poor: who and where are Colombia’s less well-off? ................................................... 8

Welfare dynamics: socioeconomic mobility of the Poor, Vulnerable and Middle-Class ....... 111

Shared Prosperity ...................................................................................................................................... 133

Inequality .................................................................................................................................................... 155

II. The drivers behind the changes in poverty and inequality .......................................................... 177

Understanding drivers of poverty reduction ........................................................................................ 177

Understanding the drivers of inequality changes ............................................................................... 211

Annex 1: Informality trends .......................................................................................................................... 233

Annex 2: Profile of the Poor, Vulnerable and Middle Class .................................................................. 244

Annex 3: Profile of the urban and rural population ................................................................................ 245

2

I. Poverty, Inequality and Shared Prosperity

Poverty: recent trends

Colombia’s solid economic growth since early 2000s has led to significant social

improvements. Since the turn of the century, extreme poverty in Colombia almost halved,

falling from 17.7 percent in 2002 to 7.4 percent in 2017. Similarly, moderate poverty fell

from 49.7 percent to 26.9 percent over the same period (Figures 1 and 2). In absolute

terms, the number of poor individuals in Colombia declined from about 20 million in 2002

to approximately 12.8 million in 2017.

The downward trend in poverty was halted in 2016, however it went back to its downward

trend on 2017. From 2016 to 2017 both moderate poverty and extreme poverty decrease

in 1.1 percentage points (p.p.), moderate poverty went from 28 to 26.9 percent, while,

extreme poverty was 8.5 percent in 2016 and 7.4 in 2017. Such decrease was primarily

driven by a lower incidence of poverty in rural areas, where extreme and moderate

poverty rates fell respectively by 2.7 and 2.6 p.p. Similarly, the urban areas saw a

reduction of moderate poverty (from 24.9 to 24.2 percent), while extreme poverty rate

has remained virtually flat since 2014, at around 5 percent.

Figure 1: Extreme Poverty Figure 2: Moderate Poverty

Source: World Bank staff calculations based on GEIH, DANE

Note: Poverty estimates are based on official poverty lines. Given the methodological changes that took place in 2006

and 2007, only the statistics reported for the 2002-05 and 2008-17 periods are comparable.

17.7

13.8

16.4

10.4

7.4

33.1

27.8

32.6

22.8

15.4

12.2

9.111.2

6.65.0

0

5

10

15

20

25

30

35

20

02

20

03

20

04

20

05

20

06

20

07

20

08

20

09

20

10

20

11

20

12

20

13

20

14

20

15

20

16

20

17

Pe

rce

nta

ge

(%

)

National Rural Urban

49.7

45.042.0

32.7

26.9

61.7

56.4 56.6

46.8

36.045.5

41.137.4

28.424.2

0

10

20

30

40

50

60

70

20

02

20

03

20

04

20

05

20

06

20

07

20

08

20

09

20

10

20

11

20

12

20

13

20

14

20

15

20

16

20

17

Pe

rce

nta

ge

(%

)


3

Colombia’s poverty reduction has mirrored that of the Latin America and Caribbean

region (LAC) and the levels have been converging in the latest years (Figure 3). In 2002,

Colombia’s poverty rate was 52.3 percent, about 7 p.p. higher than LAC, measured as

the share of the population living on less than US$ 5.5 a day (2011 PPP). This gap shrank

by less than 3 p.p. by 2015. Up to 2010, both Colombia and LAC reduced poverty at a

similar pace (-1.7 percent annually). However, in most recent years, Colombia’s pace

speeded up. After 2012 Colombia’s poverty fell at rate of -5.5 percent annually,

contrasting that of the rest of the region (-2.0). Even though poverty levels are still high

compared to other countries in the region, Colombia was among the countries in LAC

that cut poverty by the largest amount since the early 2000s (over 30 p.p.). 1

Figure 3: Evolution of poverty in Colombia and LAC

Source: World Bank calculations based on the Socio-Economic Database for Latin America and the

Caribbean (CEDLAS and The World Bank). Poverty is defined as the percentage of

individuals living on less than $5.5 US dollars a day (2011 PPP). GDP per capita growth from

World Development Indicators, The World Bank.

At the same time, Colombia achieved a significant reduction in its official

Multidimensional Poverty Index (MPI). The official MPI fell from 49 percent in 2003 to 17.0

percent in 2017 (see Box 1). This implies that 5.4 million Colombians have overcome the

multidimensional poverty threshold.

1 By 2014 Colombia ranked 12th out of 18 Latin American and Caribbean countries for which data is available.

0

2

4

6

8

10

12

0

10

20

30

40

50

60

20

02

20

03

20

04

20

05

20

08

20

09

20

10

20

11

20

12

20

13

20

14

20

15

GD

P p

c g

row

th (

%)

Po

ve

rty r

ate

(%

)

GDP pc growth COL Colombia Poverty LAC Poverty

4

Box 1: Multidimensional Poverty in Colombia

In Colombia, the Multidimensional Poverty Index (MPI) was adopted as a monitoring

instrument for public policy based on the Plan Nacional de Desarrollo 2010-2013 “Prosperidad

para todos”. Furthermore, in 2012, the CONPES 150 introduced the MPI as an official poverty

measure, that would complement the existing monetary measures. The selection of indicators

that compose the current MPI was based on data available through the Encuesta de Calidad

de Vida (ECV), and their relationship with the national development plan.

Since 2010, Colombia has been estimated the MPI following the methodology developed by

Alkire and Foster (2007). Currently, the index is comprised of five dimensions (education, youth

and children, labor, health and housing), each of them with several indicators. The official MPI

fell from 30.4 percent in 2010 to 17.0 percent in 2017 (Figure B1), implying that over 5 million

Colombians are today considered non-poor in multidimensional terms.

Figure B1.1 Multidimensional Poverty Reduction

Source: World Bank calculations based on data of the Departamento Administrativo Nacional de Estadística

(Colombia’s National Administrative Department of Statistics), DANE

Colombia has experienced an improvement in 14 of the 15 indicators that composed the MPI

(Figure B2). The main drivers of the official Multidimensional Poverty reduction have been the

increase of educational attainment, the increase in the access to the health system and the

reduction of the informal employment. Regarding the health dimension, an additional 10.7%

of the population got access to the health system in the last 8 years mainly due to the

expansion of the subsidized regime (SR), particularly after 2004.

Though a progress regarding informal employment has been observed, the share of people

working informally is still high (62.36% of Colombian in 2017).2 Long-term unemployment is the

2 World Bank estimates based on GEIH. Informality is defined as the share of total workers not contributing to

retirement funds. See Annex 1 for informality trends based on this definition since 2008, by geographic area.

0.0

10.0

20.0

30.0

40.0

2010 2011 2012 2013 2014 2015 2016 2017

Pe

rce

nta

ge

(%

)

5

only indicator that presents a deterioration, especially during the period 2012-2017. These

results are related to the weaker economic performance that the country faced after 2014.3

More importantly, out of the 15 indicators considered in Colombia´s MPI, universal coverage

has almost been achieved on 4 of them: adequate walls, adequate floors, non-child labor

and school attendance; with only a small portion of private households remaining deprived

of these opportunities (2.2%, 3.7%, 2.7% and 3.2%, respectively). As a result, the Colombian

government has decided to revisit and update the cutoffs and weights of the panel of

indicators that compose the current MPI. The new measure should consider the improvements

of welfare that Colombia had experienced in recent years.

Figure B1.2. Evolution of Colombia’s Multidimensional Poverty by indicator, 2002-2012-2017

Source: World Bank calculations based on data from the Departamento Administrativo Nacional de Estadística

(Colombia’s National Administrative Department of Statistics), DANE

3 In recent years Colombia has experienced macroeconomic shocks that include a drop-in oil prices and

the devaluation of the peso.

-14 -12 -10 -8 -6 -4 -2 0 2 4

Low Educational Achievement

No Health Insurance

Informal Employment

Critical Overcrowding

Educational Gap

Illiteracy

Poor Access to Childcare Services

No Access to Drinking Water

No Adequate Floors

No Acess to Sanitation

Child Labour

Low School Attendance

No Adequate External Walls

No Access to Healthcare Services when needed

Long-Term Unemployment

Percentage points

Progress 2010 - 2017 Progress 2010-2012 Progress 2012-2017

6

The welfare gains in the last decade were experienced by Colombians throughout the

income distribution. Whilst monetary poverty decreased, the percentage households in

the middle class has been increasing (Figure 4). In 2013, the size of the middle-class size

matched the share of poor, and for the first time in history, in 2014 it surpassed it. Currently,

less than one third of Colombian households lives in poverty and about a third is classified

as middle class. The vulnerable group surpassed poverty earlier in 2011 and today is the

largest socioeconomic group (about 41%). In other words, over 18.5 million Colombians

remain vulnerable to falling back into poverty in 2017. In spite the progress achieved in

earlier periods, the socioeconomic structure of Colombia has not changed much since

2014.

Figure 4: Evolution of poverty, vulnerable and middle class

Source: Tabulations of SEDLAC database (CEDLAS and The World Bank)

Note: Poor are individuals are considered as those living on per capita incomes less than U$S 5.3 a day (2011 PPP),

vulnerable are those with per capita incomes between US$ 5.3 and U$S 13.1 a day (2011 PPP), and the

middle and upper class as those with per capita incomes higher than US$13.1 a day (2011 PPP).

43.0 37.4

30.8 26.934.1

36.3

38.4 40.8

22.926.3

32.3

0

5

10

15

20

25

30

35

40

45

2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Per

cen

tage

Poor Vulnerable Middle + Upper Class

7

Box I : The Growth of Colombia’s Middle-Class vs other Latin American Countries

Over the last 13 years, the Latin America’s middle class grew at a faster pace than in the

1990s. Despite significant variation across countries, the region’s middle classed experienced

overall positive growth from 2002 to 2015 (see Figure B2.1). For the first time, in 2011 the LAC

region had more people in the middle class than in poverty (World Bank 20134). The trends of

declining poverty and a growing middle class were observed up to 2015.

Colombia performed similar than some of its neighbors in the Andean region — e.g., Bolivia

and Ecuador, which had comparable middle-class populations in 2002. However, several

countries outperformed Colombia transitioning the poor and the vulnerable into the middle

class —e.g., Uruguay, Argentina, Costa Rica, and Chile. As of 2015, Colombia had the

seventh smallest middle class in LAC.

Figure B2.1 The Rise of Middle Class in LAC (Circa 2002-15)

Source: World Bank calculations based on the Socio-Economic Database for Latin America and the Caribbean

(CEDLAS and The World Bank).

Note: Circa 2002 denotes other years for Argentina (2004), Chile (2003), Guatemala (2000), Ecuador (2003), Paraguay

(2003), and Peru (2004); circa 2012 denotes other years for Chile (2011), the Dominic Republic (2011),

Guatemala (2011), Honduras (2011), and Paraguay (2011). The definition of middle class is based on Ferreira

et al. (2013).

4World Bank, 2013. “Shifting Gears to Accelerate Shared Prosperity in Latin America and the Caribbean”.

Latin America and the Caribbean Poverty and Labor Brief. World Bank, Washington, DC. Disponível em.

0

10

20

30

40

50

60

70

Uru

gu

ay -

urb

an

Ch

ile

Arg

en

tin

a -

urb

an

Co

sta

Ric

a

Bra

zil

Pa

rag

ua

y

Pe

ru

Bo

livia

Co

lom

bia

Ec

ua

do

r

Do

min

ica

n R

ep

ub

lic

Me

xic

o

El S

alv

ad

or

Ho

nd

ura

s

Gu

ate

ma

la

LAC

Pe

rce

nta

ge

Circa 2002 Circa 2015

8

Profiling the poor: who and where are Colombia’s less well-off?

Colombia’s poor display more precarious indicators than other socioeconomic groups.

Compared to the vulnerable and middle-class populations, the poor in Colombia are

characterized by larger households, lower levels of educational attainment, less likely to

work, and more likely to be unemployed. Poorer households also have higher

dependency ratios due to higher numbers of children under the age of 12. In 2017, the

dependency ratio for the poor was more than twice that of the middle class.5

Despite the outstanding declines in the incidence of poverty at a national level over the

last fourteen years, historical disparities across geographical areas persist. In 2017,

extreme poverty was over 3 times higher in rural areas than in urban, and moderate

poverty was 55% higher (Figures 1 and 2). Poverty reduction in rural areas has also been

slower. Between 2002 and 2017, both extreme and moderate poverty in rural areas fell

by 5 and 3.5 percent annually. In urban areas, the reduction rate was 5.8 and 4 percent

annually, respectively. However, these averages hide different dynamics: while poverty

reduction was biased to urban areas during the first decade (extreme and moderate

poverty falling at an annual rate of 6 and 5 percent) making urban settings more

effective at lifting Colombians out of poverty, rural areas lead the reduction after 2012.

During 2012-2017, the annual rates of poverty reduction in rural areas were around 2

percentage points higher than in urban areas. This reduced the rural-urban poverty

headcount ratio from 1.65 to 1.48 percent in five years.

Sub-nationally, the incidence of poverty varies considerably even among urban areas.

Table 1 compares poverty rates in the 13 main metropolitan areas with rural and other

urban areas (Table 1). Between 2002 and 2012, the share of poor Colombians living in

main urban areas as well as those residing small and medium urban areas, decreased

both by about 17 percentage points (p.p). However, the pace of poverty reduction in

the main 13 urban areas was significantly faster (6.7 vs 3.4 percent annually). Even though

poverty reduction in small and medium areas was a bit higher (5.7 p.p.) than main urban

areas (3.2 p.p.) between 2012 and 2017, the speed of poverty reduction in the latter was

again faster. By 2017, the poverty rate in small and medium urban areas stood 2.4 times

higher than in main metropolitan areas (36.5 vs 15.7 percent, respectively). In fact, the

rate in small and medium urban areas is currently above rural marks (36.0 percent).

5 Annex 2 contains detailed profiles of the three socioeconomic groups in 2002 and 2017.

9

Table 1: Poverty Incidence Across Areas

Extreme Poor

Headcount ratio

2002 2008 2012 2017

Rural 33.1% 32.6% 22.8% 15.4%

Other urban areas 19.2% 19.5% 11.4% 8.4%

Main urban areas (13 A.M) 7.6% 5.6% 3.3% 2.7%

Millions of habitants

Rural 3.4 M 3.4 M 2.4 M 1.7 M

Other urban areas 2.3 M 2.6 M 1.6 M 1.3 M

Main urban areas (13 A.M) 1.3 M 1.1 M 0.7 M 0.6 M

Moderate Poor

Headcount ratio

Rural 61.7% 56.6% 46.8% 36.0%

Other urban areas 59.2% 52.6% 42.2% 36.5%

Main urban areas (13 A.M) 36.2% 27.0% 18.9% 15.7%

Millions of habitants

Rural 6.4 M 5.9 M 5.0 M 3.9 M

Other urban areas 7.1 M 7.0 M 5.9 M 5.5 M

Main urban areas (13 A.M) 6.5 M 5.2 M 3.9 M 3.4 M

Source: World Bank calculations based on the Socio-Economic Database for Latin America and the Caribbean (CEDLAS

and The World Bank).

Between 2002 and 2017 poverty fell across departments, but the gap among them persist.

In addition to the large urban-rural gaps, poverty data at the departmento level unveils

deeper geographical disparities, one of the most important challenges Colombia has on

fostering a balanced territorial development. Colombia could reduce the gap across

departamental poverty rates by 2017, if and only if:

I. The Departamentos that had a higher poverty headcount than the national in 2002,

experienced a faster annual rate of poverty reduction than the national between

2002 and 2017; and

II. The Departamentos that had a lower poverty headcount than the national in 2002,

experienced a lower annual rate of poverty reduction than the national between

2002 and 2017.

Figure 5 shows the poverty headcount, aggregate and by Departamento, on 2002 and

2017. It also shows the average annual rate of poverty reduction by Departamento and

the national one.

10

Figure 5: Moderate Poverty and Annual Poverty Reduction Rates between 2002-2017,

by Departamento

Source: World Bank calculations based on GEIH, MESEP - DANE.

Over the last fourteen years, Colombia did not experienced convergence in poverty

headcounts across Departamentos. 11 of the 16 Departamentos that had higher poverty

rates than the national in 2002, diverged from the national mean in 2017, since their

annual rates of poverty reduction were lower than the national one. For instance, in 2002

Chocó’s moderate poverty rate was 18 p.p. higher than the national rate. While national

poverty fell at an average annual rate of 4 percent between 2002 and 2017, for Chocó

the speed was less than 1 percent per year. The combination of these two facts leads to

an increase of the distance of Chocó’s moderate poverty relative to the national rate,

which by 2017 was more than double (about 59 percent). On the other hand, 4 of the 8

Departamentos in which poverty was lower than the national rate in 2002 experienced

a faster poverty reduction than the one seen nationally between 2002 and 2017. In the

case of Santander, in 2002 moderate poverty was 0.9 times national levels (45 and 49.7

percent, respectively). By 2017 that fraction became 0.7(18.9 vs 27 percent, respectively).

At 5.6 percent annually, Santander experienced one of the fastest poverty reduction

rates over the period of analysis.

0

1

2

3

4

5

6

7

8

9

0

10

20

30

40

50

60

70

80

Na

tio

na

l

Hu

ila

Su

cre

Ch

oc

ó

La G

ua

jira

Bo

ya

cá

Na

riñ

o

Có

rdo

ba

Ma

gd

ale

na

Bo

líva

r

Ce

sar

Ca

uc

a

No

rte

de

Sa

nta

nd

er

Tolim

a

Ca

qu

etá

Cu

nd

ina

ma

rca

Atlá

ntic

o

An

tio

qu

ia

Qu

ind

ío

Ca

lda

s

Sa

nta

nd

er

Me

ta

Va

lle d

el C

au

ca

Ris

ara

lda

Bo

go

tá D

.C.

Pe

rce

nta

ge

(%

)

Poverty in 2002 Poverty in 2017 Annual rate of poverty reduction

11

Welfare dynamics: socioeconomic mobility of the Poor, Vulnerable and Middle-Class

More than half of the Colombians remained in the same income category between 2008

and 2016. Balcazar, et al. (2017) shows that roughly 60 percent of the population remains

in the same income category (i.e. poverty, vulnerable or middle-class). However, the

vulnerable is the most mobile group among them. More than 60 percent of vulnerable

household heads shows either movements up or down. Moreover, the vulnerable

Colombian is mainly an upwardly mobile group, irrespective of the pair of years

considered in the analysis. For instance, between the initial and end year on average,

more than one third of the vulnerable group improves its income status to middle class

while around one-fourth falls back into poverty. 6

Female and male headed households experienced similar socioeconomic mobility in

the past decade. Even though poverty rates among female household heads are higher

(i.e. about 4 percentage points) than their male counterparts, welfare dynamics are

similar across these categories and over time (Balcazar, et al. (2017)). While poverty

shrank by 15.1%, 9.4%, 6.3% and 3% for male household heads in each of the four periods,

for females it did by 13.6%, 9.2%, 6.7% and 3%, respectively. Yet, female household heads

are slightly less likely to escape poverty in every period than their male counterparts

(Figure 6).

Households headed by young, uneducated or female heads had lower chances to

escape poverty during the 2008-2016 period. Figure 6 shows the rates at which poor

individuals escaped poverty and moved to vulnerability in four different time periods for

the overall poor population and different observable characteristics of the head of the

household such as gender, educational levels and age. The rates to escape poverty and

join the vulnerable segment are slightly higher among household headed by males than

by females. More importantly, these rates increase along with the level of education of

the household head. Most critically, households with uneducated heads are substantially

less likely to move up the ladder than any other education groups. Interestingly, the

highest mobility rates observed occurred in the 2010-2012 interval. Lastly, for households

that fell into poverty, the results mirror those mentioned above (i.e. the likelihood to fall

into poverty is higher for households with less educated, female and younger heads).

6 Balcazar, Dang, Malasquez, Olivieri and Pico (2018) “Welfare Dynamics in Colombia: Results from synthetic

panel”, Policy Research working paper; no. WPS 8441

12

Figure 6. Rates of Escaping Poverty to Vulnerability

(Periods 2008-10, 2010-12, 2012-14 and 2014-16)

Source: Balcazar, Dang, Malasquez, Olivieri and Pico (2017)

Note: Poor are individuals living in households for which their per capita income is less than the official poverty

lines (i.e. U$S 5.3 a day 2011 PPP), vulnerable are those households with per capita income are between

the official poverty line and the vulnerability line of U$S 13.1 a-day in 2011 PPP and middle-upper class

those with income per capita greater than the vulnerability line.

13

Shared Prosperity

Along with poverty reduction, growth in Colombia has been pro-poor as the incomes of

the poorest 40 percent of the population have grown above the national average.7

Between 2002 and 2017, Colombian households in the bottom 40 percent of the income

distribution experienced a growth in their per capita incomes of 4.0 percent per year,

compared to 2.5 for the entire population. The shared prosperity premium -the additional

growth experienced by the bottom 40 relative to the mean- is positive for the whole

period of study and across geographical areas. Apart from rural areas, shared prosperity

at the national and urban levels were evidently higher during the first decade, and

slowed down in the last five years. These lower shared prosperity outcomes have also

been experienced in several Latin American countries, because of the economic

slowdown that affected the region in the last few years.

Figure 7. Shared Prosperity, by geographical areas

Source: World Bank calculations based on GEIH MESEP-DANE

Shared prosperity outcomes are heterogeneous geographically and over time. During

the first decade the incomes of both, overall and bottom 40 population, grew more in

urban areas than in rural areas (i.e. 4.3% and 3.3% in urban, vis-à-vis 3.8% and 2.3% in rural

areas, Figure 7). Nevertheless, during the period 2012-17, rural areas experienced a larger

growth rate for both the overall population and those in bottom 40 percent (5.3% and

4.2%, respectively, vis-à-vis 2.5% and 0.5% in urban settings). Though urban areas had

7 The World Bank’s Shared Prosperity Index (SPI) measures the annualized growth rate of the average welfare

measure among the bottom 40 percent of the population. A comparison with the income growth of the

overall population sheds light on how the gains of economic growth are shared across society (World bank,

2016).

0

1

2

3

4

5

6

2002 - 2012 2012 - 2017 2002 - 2017 2002 - 2012 2012 - 2017 2002 - 2017 2002 - 2012 2012 - 2017 2002 - 2017


An

nu

aliz

ed

inc

om

e g

row

th ra

te (

%)

All Bottom 40%

14

lower mean income growths in the last five years, the share prosperity premium was larger

than in rural areas, 1.4 and 1.3 percentage points, respectively.

Over the same period, department-level improvements in shared prosperity were robust

(Figure 8). Between 2002 and 2017, shared prosperity -measured as the annualized

income growth of the bottom 40- was generally higher than the overall average income

growth across departments, resulting in a narrowing of the income gap between the less

well-off and the average Colombian. As observed in the graph, the incomes of the

bottom 40 grew up faster in most departamentos than the average income during the

2012-2017 period, if compared to the 2002-2012 one. On the one hand, during 2002-2012,

most of the less populated departments displayed a larger growth of their average mean

incomes than of the bottom 40 (those below the dotted gray line). On the other hand,

during 2012-2017 the shared prosperity premium for those departamentos was larger

than the national mean; for example, the Cordoba’s share prosperity premium was 6

percent, while Bogota’s premium was around 0 percent.

Figure 8: Shared Prosperity Improvements at department-level

Source: World Bank calculations based on GEIH MESEP-DANE

ANT

ATL

BOG

BOY

CAQ

CAU

CES

CUN

CHO

HUI

LAG

MAG

MET

NAR

NSARIS

SANSUC

ANT

BOG

BOL BOY

CAU

CES

COR

CHO

HUI

LAG

MET

NAR

QUI

RIS

SUC

TOL

-5%

-3%

-1%

1%

3%

5%

7%

9%

11%

-1% 0% 1% 2% 3% 4% 5% 6% 7% 8% 9% 10% 11%

An

nu

alize

d G

row

th r

ate

of b

ott

om

40

%

Annualized Growth rate mean income

2002-2012 2012-2017 Linear (2002-2012)

Linear (2012-2017) Linear (45 line)

15

Inequality

The higher income growth rates for those at the bottom of the income distribution have

translated into lower inequality levels. Colombia’s Gini coefficient declined from 57.2 in

2002 to 50.8 in 2017. However, inequality in Colombia remains higher than the region’s

average (LAC’s Gini was 50.5 in 2015). The growth patterns of real per capita incomes

during 2002-2017 are consistent with the inequality measures (Table 2). For example, the

relative differences in per capita income levels between the richest 10 percent and the

bottom 10 percent of the income distribution declined from 13.4 percent in 2002 to 10.0

percent in 2017, with the biggest decline happening in the 2012-17 period. However, the

richest 75 percent and the bottom 25 percent of the income distribution remained

virtually unchanged between 2002 and 2012 at about 3.6 percent, but declined by 11

percent in the following period.

Table 2. Inequality measure

2002 2012 2017

Gini 0.572 0.539 0.508

Theil 0.691 0.580 0.511

p90/p10 13.351 12.197 9.954

p75/p25 3.616 3.613 3.244

Annualized changes

Gini -0.60 -0.58

Thei -1.74 -1.26

p90/p10 -0.90 -2.01

p75/p25 -0.01 -1.07

Source: World Bank estimates based on GEIH,

DANE.

Furthermore, the Gini coefficient and the Theil index present a lower declined rate

between 2002 and 2012, than during the last four years. Even though Colombia speeded

up the reduction in inequality, countries like Bolivia and Honduras with comparable or

even higher levels of inequality in 2002, achieved better results in reducing income

inequality over the 13-year span. Finally, by 2015 Colombia was the second most unequal

country in LAC the region,8 one of the most unequal regions of the world (Figure 9).

8 Ranking is based among countries for which microdata is available in that particular year. LAC Equity Lab

tabulations of SEDLAC (CEDLAS and The World Bank) and World Development Indicators (WDI).

16

Figure 9: Colombia is the second most

unequal country in LAC Figure 10: Inequality experienced the

largest decline since 2010.

Source: LAC Equity Lab tabulations based on SEDLAC (CEDLAS and World Bank) and World Development

Indicators (WDI).

Low mobility of income held by quintiles accompanies Colombia’s high and persistent

inequality. Some income redistribution took place in 2002-17 (right side of Figure 11),

primarily driven by persistent declines in the income held by the top quintile, and

increases in the income held by the third and fourth quintiles. However, the gap between

the total income held by the richest quintile and the share held by the bottom 40 percent

of the population remained wide at the end of the period. In 2017, the year with the

lowest level of inequality, the richest 20 percent of the population held about 55 percent

of total income, while the bottom 40 percent held around 12 percent of total income

(left side of Figure 11). Moreover, while the income share of the bottom 40 percent

increase over the period, it did so only marginally.

0.505 0.511

0.200

0.250

0.300

0.350

0.400

0.450

0.500

0.550El S

alv

ad

or

Uru

gu

ay -

urb

an

Pe

ru

Do

min

ica

n R

ep

ub

lic

Bo

livia

Ec

ua

do

r

Ch

ile

Pa

rag

ua

y

Co

sta

Ric

a

Ho

nd

ura

s

LAC

Pa

na

ma

Co

lom

bia

Bra

zil

Gin

i (2

01

5)

0.583

0.555

0.511

0.400

0.450

0.500

0.550

0.600

200

2

200

3

200

4

200

5

200

6

200

7

200

8

200

9

201

0

201

1

201

2

201

3

201

4

201

5

Gin

i co

eff

icie

nt

Bolivia Brazil Colombia

LAC Mexico Peru

Uruguay

17

Figure 11. Share of total income held by each quintile

Source: LAC Equity Lab tabulations based on SEDLAC (CEDLAS and World Bank) and World Development

Indicators (WDI).

Nevertheless, the redistribution process that took place during the last five years can be

described as “pro-poor”. Between 2012 and 2017, when Colombia experienced the

largest decline in inequality, the redistribution of income benefited the poorest

population and the vulnerable class more than what it did during the 2002-2012 period

(right side of Figure 11). In conclusion, persistently high levels of inequality limited the

growth’s effect on poverty reduction.

II. The drivers behind the changes in poverty and inequality

Understanding drivers of poverty reduction

In the last fourteen years, Colombia experienced outstanding poverty reduction, and the

decline was larger in urban areas. As reviewed earlier, between 2002 and 2017 moderate

poverty in Colombia was cut by more than 40 percent (from about 50 percent to 27

percent). In urban areas the reduction was of 46 percent (from 45.5 to 24.2 percent),

while in rural of 41 percent (from 61.7 to 36 percent, over the same period). Throughout

the analyzed period and across geographical areas, poverty reduction is primarily

explained by higher per capita incomes, and in a lesser extent by better redistribution.

Figure 12 presents the results of the Datt-Ravallion decomposition of moderate and

extreme poverty measures for the country and areas over time.

2.9

6.6

10.8

18.1

61.6

3.2

7.2

11.8

19.6

58.2

3.8

8.0

12.6

20.1

55.5

0.0 20.0 40.0 60.0 80.0

Poorest

quintile

Q 2

Q 3

Q 4

Richest

quintile

Percentage

2017 2012 2002

0.91.4

1.8 2.0

-6.1-7.0

-6.0

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

2.0

Poorest

quintile

Q 2 Q 3 Q 4 Richest

quintile

Pe

rce

nta

ge

po

ints

2002 -2012 2012-2017

18

Figure 12: Drivers of poverty change: household income growth and distribution

Moderate Poverty

Extreme Poverty

Source: World Bank calculations based on GEIH, DANE.

Going in-depth, poverty reduction can be decomposed into different income

components. The observed changes in poverty can be attributed to changes in the share

of occupied adults (access to labor markets), changes in labor income (rewards and

distribution of skills), and changes in non-labor income components such as transfers

(public and private), pensions, capital, housing and others. Understanding the relative

importance and the dynamics of each of these factors may help to shed some light on

the main drivers of poverty changes. Figure 13 shows the decomposition of total poverty

into the main components of the household income that can help explain it.

-1.6 -2.7 -3.1 -3.9 -3.3 -0.4-5.3 -5.9

-3.3

-15.3 -14.4 -11.8

-1.9-0.8

-10.4

-17.4 -15.3-22.5

-30.0

-25.0

-20.0

-15.0

-10.0

-5.0

0.0

Nation Urban Rural Nation Urban Rural Nation Urban Rural

2002 - 2012 2012-2017 2002-2017

Redistribution Growth

0.1

-0.2 -1.6 -2.2-1.3 -0.6

-2.5 -1.8 -1.5

-7.3-5.5

-8.7

-0.8-0.2

-6.7

-7.8

-5.4

-16.3

-18.0

-16.0

-14.0

-12.0

-10.0

-8.0

-6.0

-4.0

-2.0

0.0

2.0

Nation Urban Rural Nation Urban Rural Nation Urban Rural

2002 - 2012 2012-2017 2002-2017

Redistribution Growth

19

Figure 13: Decomposition of changes in poverty, by income sources (2002-2017)

Moderate Poverty

Extreme Poverty

Source: World Bank calculations based on GEIH, DANE

In Colombia, moderate and extreme poverty reduction was primarily driven by labor

market over the last fifteen years. The labor market effects include changes in labor

income, changes in employment status of household members or a combination of both.

More than 70 percent of the total change in moderate poverty is related to labor market

components where 46 percent was due to labor income and 27 percent to employment.

Similar behaviors were observed in poverty reduction in urban and rural areas. However,

-6.2

-10.6

-3.4

-0.7 -0.6 -1.2

-6.3

-10.0

-2.7

-0.6 -0.5-1.2

-6.1

-15.7

-4.2

0.6 0.4

-0.7

-18.0

-16.0

-14.0

-12.0

-10.0

-8.0

-6.0

-4.0

-2.0

0.0

2.0

Share of

occupied

Labor income Transfer Pension Capital Housing+Other

non labor

income

Labor income Non-Labor income

Nation Urban Rural

-2.1

-5.0

-3.2

0.1

-0.1

0.0

-1.7

-3.7

-2.3

0.2 0.0 0.3

-3.3

-11.9

-4.6

1.0 0.80.2

-13.0

-11.0

-9.0

-7.0

-5.0

-3.0

-1.0

1.0

Share of

occupied


non labor

income


Nation Urban Rural

20

the magnitudes of the contributions were significantly higher in rural areas where 85

percent of the total reduction in rural poverty was driven mainly by labor income (61

percent) followed by employment (24 percent).

Public and private transfers were the most important factors among the non-labor income

components in poverty reduction between 2002 and 2017. Transfers account for 15

percent of the reduction in moderate poverty and 31 percent of reduction in extreme

poverty. These contributions were particularly higher in rural areas where they explained

almost 16 percent of the total reduction in moderate poverty in the last fifteen years.

Even though transfers played a significant role in poverty reduction between 2002 and

2017, these lost their leading role in reducing poverty in urban areas among the non-

labor income factors in the last five years. The geographical dynamics are not entirely

similar when we account for time. The fourteen-year span can be divided into two

periods: first decade (2002-2012) coincides with a relatively faster expansion of the

economy; and the second period (2012-2017) coincides with a relatively lower economic

growth with devaluation, plunge of oil prices, inflation and increase in unemployment,

particularly in urban areas. While decomposition results for the first decade are similar to

those presented for the fourteen-year period (i.e. labor income is the main factor),

transfers contributed to the reduction in moderate poverty rural areas (i.e. 20 percent)

but not in urban areas. Nevertheless, transfers were important in reducing the distance

to the poverty line in both areas.

Figure 14: Decomposition of changes in poverty, by income sources (2012-2017)

Moderate Poverty

-1.4

-3.8

-0.4-0.1

0.2

-0.3

-1.1

-3.1

0.0 0.0

0.2

-0.1

-2.3

-6.6

-1.6

0.1 0.2

-0.7

-7.0

-6.0

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

Share of

occupied


non labor

income

Labor income Non-Labor incomeNation Urban Rural

21

Extreme Poverty

Source: World Bank calculations based on GEIH, DANE

Understanding the drivers of inequality changes

Labor income was also the main driver behind inequality reduction in Colombia, followed

by transfers. As seen in Figure 15, labor income accounted for 74 percent of the observed

reduction in inequality in Colombia between 2002 and 2017. Below, Figure 16 illustrates

the level of inequality of each income source as measured by the pseudo-Gini

coefficient. The highly unequal distribution of labor income, coupled with the fact that

labor income represented more than two-thirds of total family income throughout the

2002–17 period, explains Colombia’s persistently high inequality. For instance, like the Gini

of total income, the pseudo-Gini corresponding to labor income declined from 0.51 in

2002 to 0.46 in 2017, primarily since 2011. Another important driver of inequality reduction

was transfers. Representing less than 5 percent of total family income, transfers are linked

to a 30.9 percent decline in inequality (Figure 15). Accordingly, the pseudo-Gini of

transfers had an outstanding reduction of 22 points, from 0.43 in 2002 to 0.21 in 2017

(Figure 16).

On the other hand, pensions had an unequal effect. Pensions, being primarily held by

those in the upper part of the income distribution, have a pseudo-Gini coefficient of

around 0.72. Despite the unequal distribution and the fact that pensions contributed to

increase inequality, it explains a small proportion of the change (i.e. 6 percent) in

inequality between 2002–2017.

-0.4

-2.0

-0.50.0

0.0

-0.1-0.2

-1.3

-0.1 -0.1

0.0 0.0

-1.0

-4.9

-1.5

0.2 0.3

-0.4

-6.0

-5.0

-4.0

-3.0

-2.0

-1.0

0.0

1.0

Share of

occupied


non labor

income


Nation Urban Rural

22

Figure 15: Decomposition of inequality by income sources

Source: World Bank staff calculations using GEIH (2002-2017).

Figure16: The Stagnation of Total Inequality is Explained by the Stagnation of Labor

Income

Source: World Bank calculations using GEIH (2002-2017).

-0.020

-0.026

-0.020

0.004

-0.0020.000

-0.017

-0.005

-0.016

0.004

0.000

0.002

-0.004

-0.020

-0.003

0.000-0.002 -0.003

-0.030

-0.025

-0.020

-0.015

-0.010

-0.005

0.000

0.005

0.010

Share of

occupied


non labor

income


Co

ntr

ibu

tio

n to

th

e c

ha

ng

e

in in

eq

ua

lity

2002-2017 2002-2012 2012-2017

0.35 0.35 0.35 0.37

0.44 0.44 0.44 0.43 0.43 0.43 0.42 0.42 0.42 0.42

0.51 0.51 0.51 0.50 0.51 0.52 0.51 0.50 0.50 0.49 0.48 0.48 0.47 0.46

0.69 0.69 0.71 0.72 0.73 0.75 0.74 0.74 0.74 0.73 0.73 0.72 0.72 0.72

0.43 0.42 0.430.40 0.41

0.360.33

0.28 0.26 0.250.22

0.200.23

0.21

20

02

20

03

20

04

20

05

20

06

20

07

20

08

20

09

20

10

20

11

20

12

20

13

20

14

20

15

20

16

20

17

P-s

eu

do

Gin

i

Capital Housing+Others Labor Income Pensions Transfers

23

Annex 1: Informality trends

Note: Informal workers are defined as those who do not contribute to pension funds.

Source: World Bank calculations based on Colombia’s GEIH 2008-2017

69.43% 70.19% 70.15% 69.94% 68.97%67.31%

65.42% 64.91%63.75%

62.36%

63.55% 64.20% 64.37% 64.25%62.97%

61.40%59.42% 58.64%

57.47% 58.11%

90.06% 90.37% 89.74% 89.53% 89.75%88.35%

87.28% 87.64%86.28%

79.52%

50%

55%

60%

65%

70%

75%

80%

85%

90%

95%

2008 2009 2010 2011 2012 2013 2014 2015 2016 2017

Nation Urban areas Rural areas

24

Annex 2: Profile of the Poor, Vulnerable and Middle Class

Poor Vulnerable Middle - class

2002 2017 2002 2017 2002 2017

Household head characteristics

Age 45.5 46.0 47.5 47.5 48.7 50.2

Female household head (%) 23.5 39.7 27.1 33.7 28.1 35.5

Education

Years of education 4.7 6.0 7.0 7.0 11.0 10.5

Non- educated 15.9 12.9 7.4 8.1 2.3 2.3

Education level: Basic Primary 54.8 43.7 41.2 39.6 19.0 21.5

Education level: Basic Secondary 15.8 14.9 18.6 14.9 12.4 12.0

Education level: Middle school 11.5 21.7 23.8 26.0 23.9 24.8

Education level: Higher

education 2.0 6.8 9.0 11.3 42.4 39.4

Household characteristics

Household size 4.7 4.2 3.9 3.6 3.0 2.6

Living in urban area 61.5 70.8 85.6 68.6 92.4 91.9

Share of member age 0-12 34.7 33.5 22.3 23.5 15.3 12.3



Share of member age 70+ 2.9 3.7 3.8 4.4 4.4 6.7

Dependency ratio 79.9 80.0 48.2 52.7 35.0 34.9

Labor market (Age 15-64)

In labor force 67.5 63.8 73.6 73.4 76.9 81.5

Female labor force 49.6 50.8 62.1 61.5 70.5 74.9

Unemployment status 13.9 11.2 10.6 7.2 6.6 4.7

Employer 5.5 1.8 3.9 3.0 9.0 5.5

Employee 52.5 24.4 59.5 48.6 64.7 64.1

Self-employed 39.0 65.9 33.5 44.2 24.4 29.0

Unpaid worker 3.0 7.9 3.1 4.2 1.9 1.4

Employment sector (%)

Primary 35.4 29.4 11.6 21.7 5.0 5.6

Manufacturing 10.9 9.7 16.6 11.5 14.3 13.1

Construction 5.8 6.9 5.3 7.0 2.8 5.2

Retail 29.6 37.4 37.5 36.3 29.5 33.2

Utilities 0.3 0.2 0.5 0.4 0.8 0.8

Services 18.0 16.2 28.4 23.0 47.6 42.1

Source: World Bank calculations based on Colombia’s national household surveys (ECH for 2002-2006 and GEIH for

2008-2017)

25

Annex 3: Profile of the urban and rural population

Urban Rural

2002 2017 2002 2017

Household head characteristics

Age 46.6 48.2 47.6 48.5

Female household head (%) 28.6 39.2 16.4 23.0

Education

Years of education 7.8 9.1 3.9 5.0

Non educated 6.7 4.3 21.3 16.1

Education level: Basic Primary 36.7 27.4 60.9 54.0

Education level: Basic Secondary 18.1 14.4 8.8 11.4

Education level: Middle school 21.7 27.6 6.0 13.4

Education level: Higher education 16.8 26.3 2.9 5.1

Household characteristics

Household size 4.0 3.3 4.4 3.5

Share of member age 0-12 26.6 21.1 31.8 27.3



Share of member age 70+ 3.4 5.0 3.4 4.7

Dependency ratio 57.2 48.8 73.5 65.8

Labor market (Age 15-64)

In labor force 47.7 52.8 44.4 47.5

Female labor force 41.1 47.0 27.9 32.9

Unemployment status 10.8 6.9 6.4 3.1

Employer 5.0 4.0 11.0 4.0

Employee 52.9 52.8 66.0 33.4

Self-employed 38.2 40.5 22.6 52.5

Unpaid worker 3.9 2.7 0.4 10.0

Employment sector (%)

Primary 7.3 4.5 63.3 63.6

Manufacturing 15.8 13.3 6.5 6.3

Construction 5.5 6.8 2.7 3.6

Retail 37.8 40.6 15.3 16.1

Utilities 0.6 0.6 0.3 0.2

Services 33.0 34.1 12.0 10.0

Source: World Bank calculations based on Colombia’s national household surveys (ECH for 2002-2006 and GEIH for

2008-2017)

Poverty and Shared prosperity in...

Documents

Transcript of Poverty and Shared prosperity in...